← Back to glossary
Browse letter H hub
Hallucinations
Confident but incorrect outputs generated by AI models.
Definition
Confident but incorrect outputs generated by AI models.
How It Works
Hallucinations helps teams build predictable AI and translation workflows by setting clear expectations for quality, consistency, and decision-making.
In production environments, this concept is applied with process controls such as human review, terminology alignment, and repeatable quality checks across multilingual content.
Key Concepts
- core principle of hallucinations
- workflow-level implementation
- terminology and quality consistency
- human validation before publication
Where It Is Used
- localisation workflows
- AI translation pipelines
- multilingual content production
- cross-referencing related concepts such as Human Evaluation